As is known, the first essential step of a process for designing and planning a radiocommunications network for mobile terminals is computing the so-called cell coverage, i.e., extent and features of a region around a radio base station where radioelectric signals received by a mobile terminal and radiating out from the radio base station cope with given requirements.
Generally, this region is the locus of points where the strength, or a quantity related thereto, of a radioelectric signal received by the mobile terminal and radiating out from the radio base station exceeds a given threshold. Such a threshold may be defined by using different criteria, the most adopted of which are detectability of a reference channel in the radioelectric signal received by the mobile terminal, and transmission error rate higher than a threshold value.
Traditionally, one of the most frequently used methods for computing cell coverage includes radially scanning the region around the radio base station along angularly equispaced radial scanning line connecting the radio base station and the point where one of the following three quantities, which, considered singularly, may be regarded as indicative of the cell coverage, is to be computed: the point strength of the radioelectric signal received by the mobile terminal, the local mean of the point strength of the radioelectric signal, and the median value of the local means of the point strength of the radioelectric signal.
The point strength is the value of the modulus (or envelope) of the radioelectric signal in a given point of the region, the dimensions of point being substantially equal to those of the physical element which is used to measure the point strength of the radioelectric signal: in this case, the mobile terminal antenna of few centimeters.
The local mean is the mean value of the point strength of the radioelectric signal within some tens of wavelengths long, which, having regard to the frequencies involved in mobile radiocommunications, results in considering paths 5 to 10 meters long or areas some tens of square meters wide.
The median value of the local means of the point strength of the radioelectric signal is a resumptive statistical value which, to guarantee a satisfying reliability (confidence) thereof, is to be computed by taking account of a congruous number of local means (10 to 20), which results in considering paths 50 to 100 meters long or areas few thousands of square meters wide.
Measuring the point strength of the radioelectric signal is not presently one of the key points in the development of radiocommunications network planning tools because of the extreme spatial variability of the radioelectric signal strength due to the “fine” structure (order of magnitude of the centimeter) of the surrounding environment.
Computing the local mean of the point strength of the radioelectric signal is, nowadays, still marginal in mobile radiocommunications network planning due to the modellization and computation complexity and to the huge amount of environmental data (cartographic database) to be processed.
This quantity is generally taken into account only during coverage computation for micro cells (cells with radio base stations arranged at few meters from the ground), which are characterized by narrow territorial extents (diameter of some hundreds of meters) and which are the minority (about 10%) of the cells forming a typical mobile radiocommunications network.
Computation of this quantity is indeed hard to carry out for traditional cells, i.e. cells having a coverage area with a diameter of several kilometers, such as large cells (cells with radio base stations arranged on isolated masts) or small cells (cells with radio base stations arranged on building roofs), due to the high computation time and, above all, to the low reliability of the results at these distances.
On the contrary, computing the median value of the local means of the point strength of the radioelectric signal plays a paramount role in the development of radiocommunications network planning tools because in most cases this quantity represents the physical parameter associated with the concept of cell coverage. Therefore, nowadays the design of a mobile radiocommunications network is substantially based on a electromagnetic coverage prediction based on the median values of the local means.
A traditional cell coverage computation, sometimes also referred to as cell coverage prediction, is carried out by using a low environment resolution, i.e., by taking account of data describing the features of the environment within elementary areas, generally known as pixels, having a side of 50 or 100 meters.
The median values of the local means of the point strength of the radioelectric signal along the scanning lines are computed by taking account of the power radiating out from the radio base station, a open environment propagation curve, a morphological factor, a urbanization factor and an orographical factor (diffraction on natural obstacles).
In particular, the open environment propagation curve is indicative of the radioelectric signal strength attenuation, also known as propagation loss, in a open area, i.e., an area empty of trees, buildings or architectural structures made by human beings, and is generally expressed as a semi-empiric relation as a function of the power radiating from the radio base station antenna, antenna radiation pattern, distance from the radio base station and the mobile terminal, antenna tilt, radiation frequency, and effective antenna height with respect to the ground.
The orographical factor, urbanization factor and the morphological factor are correction factors for the open environment propagation curve and describe, respectively, the altitude features of the geographical area, the building features within areas having selectable sizes (e.g. 50 by 50 meters), and the geographical area in terms of morphological classes (forested, lakeside, etc.), which, as is known, influence radioelectric signal propagation.
In some cases, only morphological and urbanization features within the pixel, for which computation of the local means of the radioelectric signal strength is carried out, are taken into account, whereas in other cases, characterized by a more refined approach, these features all along the scanning lines are taken into account.
Lastly, computing diffraction on natural obstacles is the aspect that requires a more complex processing of the cartographic data. Starting from the orography, an altimetric profile is determined all along each scanning line, and the interaction effects (signal attenuation) with the possible natural obstacles arranged along the scanning lines are computed by resorting to the classic Huyghens-Fresnel theory, according to which such interaction effects may be assessed with adequate reliability by replacing each natural obstacle with an equivalent virtual obstacle (screen) having a knife edge shape, a height equal to the natural obstacle, an infinitesimal thickness, endlessly extending perpendicularly to the propagation direction, and perfectly absorbing the incident electromagnetic signal.
Other more advanced approaches are provided, instead, for computing diffraction on natural obstacles by resorting, rather than to an infinitesimal thick screen, to a finite thick screen having a rounded edge. However, the approaches based on an infinitesimal thick screen are the most frequently used because they are simple and are adapted to this specific problem on the basis of some known algorithms which have been derived from the literature and appropriately modified and optimized to take into account the effects due to multiple obstacles. Examples of such known algorithms are the Epstein-Peterson method, the Deygout method and the stretched string method, the latter being recommended by ITU-R 526 and being the best trade-off between result reliability and algorithm computational speed and thus the most frequently used.
For a more detailed discussion of the low environment resolution cell coverage computation, the reader is referred to the following publications, which are incorporated herein by reference, in their entirety:    1) M. Hata, “Empirical formula for propagation loss in land mobile services”, IEEE Trans. On Vehicular Technology, Vol. 29, 1980;    2) E. Damosso, L. Stola, “Radiopropagazione”, Scuola Superiore Guglielmo Reiss Romoli, L'Aquila, 1992;    3) ITU-Reccommendations Rec. P. 526-3 “Propagation by diffraction”;     4) G. Bussolino, R. Lanzo, M. Perucca, “Rasputin: a field strength prediction model for large and small cell mobile system using territorial data base”, 7th International Network Planning Symposium, Sidney 1996;    5) COST 235 “Radiowave propagation effects on next generation fixed service terrestrial telecommunication systems”, Chap. 4, Final Report EUR 16992 EN, 1996.
The need for a growing number of radio base stations, together with the need for more complete services, in particular services having features more and more detailed on a territorial level, has forced second and third generation mobile radiocommunications network providers to resort to a high environment resolution network design which enables definition of specific design parameters for very narrow territory elements. For example, along a road or in a square given services may be provided rather than others and, in any case, services with a level appropriate to the specific territorial, social and town reality.
The low environment resolution, which is typical of traditional mobile radiocommunications network design and planning, is evidently inadequate to cope with the above-mentioned needs, which may instead be partially satisfied by using a high environment resolution, i.e., by taking account of data describing the features of the environment within pixels having a side of 5 or 10 meters, which high environment resolution is more consistent with the dimensions of the town elements and, at the same time, allows computing the local means of the point strength of the radioelectric signal.
A number of different methodologies for computing high resolution environment cell coverage on the basis of the local means of the point strength of the radioelectric signal have been proposed.
For a detailed discussion of these methodologies, the reader is referred to the following publications, which are incorporated herein by reference, in their entirety:    1) EP-A-1 292 163, “Method for determining the values of the electromagnetic field generated by a radio base station in an urban environment”;     2) M. Perucca, M. Signetti “Small cells planning analysis of electromagnetic models from measurements at 1800 MHz”, ICAP 1997;    3) COST Action 231 “Digital mobile radio towards future generation systems”, Chap 4, Final Report EUR 18957, 1999,    4) ITU-R Rec. 1411 “Propagation data and prediction methods for the planning of short range outdoor radio communication systems and radio local area networks in a frequency range 300 MHz to 100 GHz”;     5) US-A-2001/0041565 “Method and apparatus for network planning”. 
All of these methodologies, however, have been designed and developed for short distances from the radio base station, in particular distances shorter than 1 or 2 kilometers, and consequently they involve a territorial analysis which is entirely carried out using a high environment resolution, i.e., considering pixels having a side of 5 or 10 meters.
Therefore, a paramount problem generally experienced in extending these approaches to large distances (10-20 km) is represented by the computation time and, above all, by the result reliability. In particular, once the refinement level of the computation model has been fixed, the result reliability depends mainly on the number of interactions with the surrounding environment along the scanning line which occur during computation of the local mean of the point strength of the radioelectric signal for the pixel considered. Inevitably, each interaction with the surrounding environment involves a given computation approximation and consequently a computation error which accrues during the computation.